Application of an image processing technique for early diagnosis and monitoring of glaucoma

Abstract

Early stage glaucoma can be diagnosed by finding retinal nerve fiber defects using a color image processing technique. This method is based on the reflectance of the defective part which has low values in the green and blue wavelengths. The level pixel (picture element)of reflectance is scored from 0 to 255. In early stage glaucoma, the difference of that reflectance is very small. This new method detects that small difference. The method consists of sampling and summation value of pixels in green and blue along the radial lines drawn between two circles whose centers are the same and located at the center of the eye. It was found that the small differences can be detected only by doctors with a sufficient experience. This method is better than the color enhancement. Also, the detection of cup and blood vessels has been shown to assist in automatic processing. It will be necessary to apply this method to many patients to check the of the utility of the method and develop automatic processing system in the future.